Graduate Students for Safe and Responsible AI (SAFARI)
Annalise Norling
Project summary
This project aims to create a graduate student association of graduate students who care about and have an interest in safe and responsible AI development and deployment, broadly construed. I mean broadly construed to imply that this association is, in my mind, ideally an interdisciplinary one, bringing together graduate students from different fields which each have unique perspectives and methods, to address concerns about the rapidly accelerating progress in artificial intelligence.
I also envision this association to be a space where rising academics and researchers who have a strong (or perhaps even a modest) interest in AI - but might not have it as their primary research interest - can also get involved and collaborate with others, including those who are more immersed in AI or have it as their primary research focus.
I plan to initiate the association at Indiana University Bloomington, specifically (at least at first) among the Philosophy graduate students (as well as several students who have expressed interest to me from the Biology and Creative Writing graduate programs as well), since that is where I am currently a graduate student and am therefore uniquely positioned to start things. But I then also hope that the association will continue to extend to other departments at IUB, and ideally then to other universities as well.
Project goals
Develop a network of rising professionals who share an interest in safe and responsible AI development and deployment, and keep track of the sorts of graduate students that are interested in the field.
Foster a long-term collaborative problem-solving environment for graduate students from various fields including graduate students who have not much expertise with the field, or a tangential interest in the field, as well as students with significant expertise and a direct interest in the field.
Foster long-term professional relationships between graduate students interested in safe and responsible AI, and in particular between graduate students from widely different fields, who may not otherwise meet or work collaboratively.
Develop long-term project goals once there is a sufficient number of members in the association (this might look something like: vote on what this might look like at around 40 members).
Develop a space for discussion of AI and how it might relate to other areas of inquiry/interest.
How will this funding be used?
My time in thinking about how to set up the initial association and gain membership, and then setting it up (~30 hours @ $40/hour - 1.2k)
Monthly association meetups (snacks, refreshments, and advertisements to put up on campus so that we get a good number of people to come - 2k)
Organize an association reading group (snack, advertisements, and a copy of the books that we will read together for each member committed to attending regularly - 6k)
Grad association colloquium series (invite graduate students from any university for a talk and Q&A, refreshments - 10k)
Create a student journal for interdisciplinary AI research and theorizing (50k).
Organize a conference (75k)
At a minimum, to start this project I will need 9.2k. If I do get that funding (and no more) I plan to do steps 1, 2, and 3. However, I think that this association could have more potential, so I have also included steps 4, 5, and 6 as goals. In particular, I think setting up a colloquium series (Step 4) would add a lot to the association.
What is your (team's) track record on similar projects?
There is a strong interest in an association of this type among philosophy and biology students that I know of from surveying those that I know - I anticipate ~20-30 members without additional advertising, etc.
I also personally have a tangential interest in AI, although it is not my primary interest, and have for several years now been hoping that such a group would come about.
How could this project be actively harmful?
If the association is very poorly organized, it could end up producing the opposite effect than anticipated. But I anticipate that by ensuring that the association is governed by a number of members and not only a select few that this can be avoided confidently.
What other funding is this person or project getting?
Irving and Shirley Brand Fellowship (5k). This fellowship is to support my work next year broadly. In the case that I do not receive additional funding, I will likely not utilize these funds for this project, given the difficulty of doing so with this amount only. But if I do receive additional funding, I will put this fellowship toward the same efforts.
Minimum Funding
At a minimum, 9.2k would allow me to get a substantive enough start on the association that I am confident it could at least sustain itself while I seek additional funding.
Funding Goal
Ranked in order of priority (not necessarily with respect to importance, but rather the timing of setup):
My time in thinking about how to set up the initial association and gain membership, and then setting it up (~30 hours @ $40/hour - 1.2k)
Monthly association meetups (2k)
Association reading group (6k)
Grad association colloquium series (10k)
Student journal for interdisciplinary AI research and theorizing (50k).
Conference (75k)